DISSERTATION DETERMINANTS OF HABITAT USE AND COMMUNITY STRUCTURE OF RODENTS IN NORTHERN SHORTGRASS STEPPE. Submitted by. Paul T.

DISSERTATION DETERMINANTS OF HABITAT USE AND COMMUNITY STRUCTURE OF RODENTS IN NORTHERN SHORTGRASS STEPPE Submitted by Paul T. Stapp Department of Biology In partial fulfillment of the requirements for the Degree of Doctor of Philosophy Colorado State University Fort Collins, Colorado Spring 1996

COLORADO STATE UNNERSITY 5 March 1996 WE HEREBY RECOMMEND THAT THE DISSERTATION PREPARED UNDER OUR SUPERVISION BY PAULT. STAPP ENTITLED DETERMINANTS OF HABITAT USE AND COMMUNITY STRUCTURE OF RODENTS IN NORTHERN SHORTGRASS STEPPE BE ACCEPTED AS FULFILLING IN PART REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY. Committee on Graduate Work 11

ABSTRACT OF DISSERTATION DETERMINANTS OF HABITAT USE AND COMMUNITY STRUCTURE OF RODENTS IN NORTHERN SHORTGRASS STEPPE Patterns of distribution and abundance of small mammals reflect the responses of individuals to the spatial and temporal availability of resources and abiotic conditions, as well as interactions with conspecifics and other species. I examined habitat selection of two rodents, the deer mouse (Peromyscus maniculatus) and the northern grasshopper mouse ( Onychomys leucogaster), on shortgrass steppe in north-central Colorado. Both species consume arthropods when these resources are plentiful, but grasshopper mice prey on other rodents and thus may have both competitive and predatory effects on deer mice. To examine these interactions, I conducted a removal experiment to determine the effect of grasshopper mice on microhabitat use, diet, and abundance of deer mice, and an odor-response experiment to determine whether olfactory cues mediate interactions between these species. Deer mice preferred shrubs at both individual and population levels, presumably to reduce predation risk. Mice oriented movements toward shrubs and traveled under shrubs more often than expected based on the density of shrubs on study plots. Population density also increased with increasing shrub density and aggregation. The response of mice to shrub cover was non-linear. Thresholds in the selective use of shrubs, movement patterns, and abundance occurred over a narrow range of shrub cover where shrubs were most aggregated, underscoring the importance of both shrub density and dispersion. Mice tended to accumulate in areas where their movements were most tortuous, suggesting that it is possible to generate testable predictions about patterns of abundance from individual movements. iii

In contrast, grasshopper mice showed no affinity for shrub microhabitats, and instead oriented movements towards rodent burrows and disturbances created by pocket gophers (Thomomys talpoides). Results from pitfall trapping in different microhabitat types suggested that grasshopper mice used gopher mounds and burrows because of the concentration of insect prey in these microhabitats. The abundance of these microhabitats also was a better predictor of grasshopper-mouse abundance than were broad-scale, qualitative descriptors of macrohabitat type. The significance of these microhabitats across scales demonstrates the importance of spatial and temporal availability of prey to grasshopper mice. Even though grasshopper mice and deer mice show different habitat affinities, grasshopper mice may affect the surface activity and abundance of deer mice in areas where they co-occur. Deer mice decreased in number throughout the removal experiment on both control and removal sites, but the decline was greatest on controls, where grasshopper-mouse numbers increased. No shifts in microhabitat use were detected on removal sites, but deer mice increased their use of shrubs on control sites when grasshopper mice were most abundant. Because diets of deer mice did not differ between control and removal sites during the experiment, grasshopper mice apparently influenced the behavior and populations of deer mice through predation or interference rather than resource competition. Increases in the abundance of granivorous rodents on removal sites support this conclusion, and suggest that grasshopper mice, when abundant, can impact the composition of local assemblages on shortgrass steppe. However, if deer mice actively avoid contact with grasshopper mice, it is unlikely that this interaction is mediated by olfactory cues. When presented with odors of grasshopper mice, harvest mice, and clean cotton, deer mice showed no avoidance of grasshopper-mouse odors, regardless of season, sex or reproductive condition of respondents, or history of contact with grasshopper mice. Paul Stapp Department of Biology Colorado State University Fort Collins, CO 80523 Spring 1996 iv

ACKNOWLEDGEMENTS This research would not have been possible without the generous assistance and support of a large number of individuals and institutions. I am particularly grateful for the guidance provided by my graduate committee, Drs. Lou Bjostad, John Wiens, and Bruce Wunder, and in particular, my advisor, BeaVan Horne. Each provided a unique perspective and philosophy that made all aspects of my graduate studies at CSU rewarding. Field studies of small, secretive mammals are both expensive and labor-intensive, and this work would not have been feasible without summer support provided by Bill Lauenroth and Indy Burke of the Shortgrass Steppe Long-Term Ecological Research (LTER) project. My interaction with the L TER scientists helped me professionally by providing a diverse arena for my ideas and by forcing me to appreciate more than just the mice. I am also grateful to the Department of Biology for allowing me to continue to improve my teaching skills and for travel funds, to the Graduate Degree Program in Ecology for teaching and research fellowships, and to the American Museum of Natural History and the American Society of Mammalogists for recognizing the potential contributions of my work. My association with the L TER project allowed me to hire and keep a number of excellent field assistants. Scott Persson, Fred Durrance, and Eric Simandle (1993), Bonnie Frank and the L TER field crew ( 1994 ), and Laura Higgins ( 1995) collected much of these data, under conditions that were often uncomfortable and always unpredictable. I'm particularly grateful to Bonnie, who spent cold nights on hands and knees, staring at purple grass, and unbearably-hot days staring at the same grass through a point frame. Bonnie collected most of the data for the removal experiment (Chapter 4) and I'll always appreciate her hard work and patience, both with me and her duties on the project. Finally, I owe a huge v

debt to Mark Lindquist, the L TER site manager, for making all aspects of the field work easier for me and for his friendship over the past four years. Mark's untethered enthusiasm was an enormous boost throughout my project and was responsible for the initiation and continued success of the monitoring programs (Chapter 6). Thanks, Mark, for letting me take the motorcycle during that thunderstorm the first summer; few lessons are so valuable or memorable. A number of other individuals provided logistical and technical assistance during the project. Mary Ashby and Jeff Thomas of the Agricultural Research Service helped me locate study sites and tolerated my frequent changes of plan. Bruce Wunder, Lowell McEwen, and the LTER loaned traps, Pat Ward permitted me to use his telemetry equipment, and John Wiens allowed me to use fencing and surveying equipment. Sue Vande Woude helped with animal care issues and surgery protocols and was a great collaborator on a side project. I also received helpful statistical advice from Phil Chapman and the staff of the CSU Statistics Laboratory. I am very fortunate to have had the opportunity to work in the company of students and postdocs who were not only talented scientists but also supportive and helpful friends. Of these, Bob Schooley, Brandon Bestelmeyer, and in particular, Jeff Kelly provided immeasurable assistance throughout the project and my stay at CSU. All three suffered through drafts of these or other manuscripts, and their insights and constructive comments improved my work more than they realize. Brandon and Kim With introduced me to the Pawnee; Kim's energy and work habits were a valuable example early in my research. Thanks also to Mary Beth Voltura, who was a great teaching partner and a good friend, especially during the past year. All aspects of this research benefitted greatly from the insights and suggestions of Tom Crist, who was a postdoctoral associate in our lab during the first three years of my program. Tom always found the time for statistical and design questions and openly shared his ideas, broad training, and philosophy. His presence anchored the lab and his guidance and constructive comments greatly improved the quality of my work. vi

Finally, I'd like to acknowledge the support of my family, who waited patiently for me to finish and didn't wonder too loudly what I was doing and why. More than anyone, however, I owe the successful completion of my studies to my wife, Laura, who served faithfully and tirelessly as critic, editor, audience, (yes) field assistant, and always, a source of perspective. Her uncompromising faith and confidence in me stimulated me to work harder, but also let me put it away when I really needed to. This work is dedicated to her, to Swampy, who started all this ten years ago, and to the mice of the Pawnee, for their reluctance in divulging their secrets. vii

TABLE OF CONTENTS ABSTRACT OF DISSERTATION... iii ACKNOWLEDGEMENTS... v I. OVERVIEW AND SUMMARY Framework, questions, and approach... 1 Study area and organisms... 3 Habitat selection of Peromyscus maniculatus and Onychomys leucogaster... 6 Effects of Onychomys leucogaster on Peromyscus maniculatus and other rodents... 8 Insights and unanswered questions... 11 Literature Cited... 12 Tables... 17 Figures... 18 viii

ll. RESPONSE OF DEER MICE (PEROMYSCUS MANICULATUS) TO SHRUBS: LINKING SMALL-SCALE MOVEMENTS AND THE SPATIAL DISTRIBUTION OF INDIVIDUALS Abstract... 19 Introduction... 20 Materials and Methods... 21 Results... 25 Discussion... 27 Literature Cited... 31 Tables... 34 Figures... 36 ix

ill. EFFECTS OF VEGETATION AND SUBSTRATE CHARACTERISTICS ON PREY AVAILABILITY AND HABITAT SELECTION OF NORTHERN GRASSHOPPER MICE (ONYCHOMYS LEUCOGASTER) Abstract... 39 Introduction... 40 Methods Study area... 42 Live-trapping, movements and habitat use... 43 Diet analysis... 45 Patterns of arthropod abundance... 46 Patterns of mouse abundance... 47 Results Live-trapping, movements and habitat use... 48 Diet analysis... 50 Patterns of arthropod abundance... 51 Patterns of mouse abundance... 53 Discussion... 53 Literature Cited... 58 Tables... 62 Figures... 68 X

IV. COMMUNITY STRUCTURE OF SHORTGRASS-STEPPE RODENTS: THE ROLES OF INTRAGUILD PREDATION AND COMPETITION Abstract... 72 Introduction... 73 Study area and species... 75 Methods Population size and removal experiment... 76 Microhabitat use... 77 Diet overlap... 79 Results Population size and removal experiment... 80 Microhabitat use... 83 Diet overlap... 84 Discussion... 85 Literature Cited... 90 Tables... 95 Figures... 98 xi

V. DO OLFACTORY CUES MEDIATE INTERACTIONS BETWEEN RODENTS ON NORTHERNSHORTGRASSPRADUE? Abstract... 103 Introduction... 104 Methods Study area... 105 Odor-response experiment... 106 Capture-recapture studies... 108 Results... 109 Discussion... 110 Literature Cited... 115 Tables... 119 Figures... 121 xii

VI. MONITORING STUDIES OF SMALL-MAMMAL POPULATIONS ON THE SHORTGRASS STEPPE LONG-TERM ECOLOGICAL RESEARCH SITE Abstract... 122 Introduction... 123 Methods Study area........................... 124 Nocturnal rodents... 125 Lagomorphs... 127 Density estimation... 127 Results and Discussion Nocturnal rodents... 128 Lagomorphs...... 130 Conclusions and future directions... 131 Literature Cited... 133 Tables... 136 Figures... 137 APPENDIX 1... 141 APPENDIX 2................................... 143 APPENDIX 3... 145 xiii

CHAPTER 1 OVERVIEW AND SUMMARY Framework, questions, and approach Patterns of distribution and abundance of animal populations reflect the success of individuals at finding mates, locating sufficient food and nutritional resources, and avoiding predators (Andrewartha and Birch 1954 ). The consequences of these activities are ultimately manifested in patterns of habitat use, which vary across a range of scales of measurement. At fine spatiotemporal scales, behavioral ecologists have been successful at documenting the decisions made by individual foragers in choosing microhabitats (Stephens and Krebs 1986). Elucidating the mechanisms underlying habitat selection at the population or macrohabitat scale is more complicated, for several reasons. First, the extent of spatial heterogeneity may exceed the dispersal capabilities of most individuals, so that access is restricted to all but subset of possible habitat types. Additionally, site-tenacious individuals that remain in an area regardless of current resource conditions may bias assessments of limiting factors (Wiens 1989). Finally, agonistic interactions between conspecifics may intensify with increasing density, so that subordinate individuals accumulate in poorer-quality habitats (Van Home 1982, Pulliam 1988). These complications disconnect resource abundance from population size and space use, and limit our ability to use local population density to assess the relative quality of habitats reliably (Van Home 1983). Yet, because of the relative ease of censusing populations, most tests of habitat-selection theory of free-ranging vertebrates have studied how individuals are distributed among habitats (Fretwell and Lucas 1970) and have not explicitly addressed habitat selection at both behavioral and population scales. 1

Understanding habitat selection is important to studies of vertebrate communities because habitat partitioning may determine coexistence of similar species and, as a consequence, community organization (Rose and Birney 1985, Kotler and Brown 1988, Kaufman and Kaufman 1989, Brown and Harney 1993). Investigations of the determinants of the structure of small mammal communities conducted during the 1970s and early 1980s emphasized the role of interspecific competition (Dueser et al. 1989). More recently, researchers have emphasized that patterns of microhabitat use reveal differences among species in their vulnerability to predators, so that patterns of distribution and abundance in local rodent assemblages reflect trade-offs between competitive abilities and predation risk (Kotler 1984, Sib et al. 1985, Kotler et al. 1994, Batzli and Lesieutre 1995). However, identifying the mechanisms that determine the structure of rodent communities ultimately requires an understanding of interactions among individuals to complement patterns derived from broad-scale studies of abundance. My research examined habitat selection of two nocturnal rodents and how resource distributions and interspecific interactions between these species influence patterns of distribution and abundance in semi-arid grasslands of central North America. The studies described here address two general questions: ( 1) How do individuals respond to spatiotemporal variation in resource availability and habitat characteristics, and how are these processes translated into patterns of distribution and abundance? (2) Given that species have different habitat affinities and requirements, what is the role of interspecific interactions in determining the structure of local assemblages, what is the nature of these relationships, and how are they mediated? My approach was to identify potentially important resources for each species, using existing knowledge from the literature and preliminary studies, and document the response of individuals to spatial and temporal variability in these resources. I used information gained from studies of the behavior of individuals to interpret patterns of habitat use measured at 2

population scales. With this knowledge, I conducted field experiments to examine how interactions between these species might contribute to patterns of local abundance, and to explore a possible behavioral mechanism through which these interactions might be mediated. Study area and organisms The study area was the Central Plains Experimental Range, a region of shortgrass steppe located approximately 60 km northeast of Fort Collins, Colorado. The Central Plains Experimental Range is managed by the United States Department of Agriculture (USDA) Agricultural Research Service and is the location of both the United States International Biological Program (USIIBP) Grassland Biome project (1968-1976) and the National Science Foundation Shortgrass Steppe Long-Term Ecological Research project (1982-present). Detailed descriptions of the vegetation, topography, and climate are provided in the following chapters, and Appendix 1 contains a list of the mammals species found on and adjacent to the USDA Forest Service Pawnee National Grasslands. However, a general description of shortgrass steppe is appropriate here to understand the context within which my field studies were conducted. Short grass steppe is dominated by blue grama (Bouteloua gracilis), which forms a largely continuous mat of short perennial grass, interspersed with patches of bare soil, plains prickly-pear (Opuntia polycantha) and several species of small shrubs (Artemisiafrigida, Gutierrezia sarothrae, Eriogonum effusum). These grasslands are relatively tolerant of grazing by cattle and resistant to invasion from native and exotic plants in the absence of significant disturbances to the soil (Lauenroth and Milchunas 1991). In general, the vegetation has little vertical structure, but at a broader scale, shortgrass steppe is best viewed as a mosaic of grassland and shrubland, with riparian vegetation along permanent streams. The lack of significant cover, coupled with the harsh abiotic conditions common at this location (40 49' N, 107 47' W), has presumably led many species to concentrate activity near shrubs and in association with subterranean refuges. 3

The dominant large shrub on the Central Plains Experimental Range is four-wing saltbush (Atriplex canescens), with small soapweed (Yucca glauca) present on ridges and sandy soils. Saltbush is primarily restricted to areas in and adjacent to seasonal drainages or low-lying swales, and reaches highest densities in the floodplain of Owl and Cow Creeks, where soils are loam and grasses are mostly blue grama or western wheatgrass (Pascopyron smithii). On the more coarsely-textured soils paralleling washes, saltbush and other small shrubs occur with a mixture of both short and intermediate-height grasses. The diversity and abundance of small mammals is greater on areas where saltbush is present, and in areas of mixed shrub grassland than on open prairie (Chapter 6). Smallmammal studies conducted during the mp Grassland Biome project focused exclusively on grassland areas and reports from these studies emphasized the low diversity and low population densities of shortgrass prairie compared to other North American grasslands (Grant and Birney 1979, French et al. 1976). Perhaps as a result of these findings, there have been relatively few investigations of the ecology of mammal populations on shortgrass steppe. However, my research, and the preliminary results of long-term monitoring I implemented through the Shortgrass Steppe Long-Term Ecological Research project (Chapter 6, Appendix 1 ), suggest that a broader view of shortgrass steppe as a landscape mosaic may lead to greater appreciation for the diversity and roles of the native mammalian fauna. My studies focused on the two most common species of nocturnal rodents (Muridae; Sigmodontinae) on the Central Plains Experimental Range. The deer mouse (Peromyscus maniculatus) is ubiquitous, inhabiting both pristine and human-dominated ecosystems throughout most of North America (Hall 1981 ). This small mouse (20 g) is regarded as a generalist in both diet and habitat (Baker 1968, O'Farrell 1980), although many studies conducted in open-canopy landscapes have demonstrated an affinity for microhabitats with vertical cover (e.g., Holbrook 1979, Thompson 1982, Travers et al. 1988). The affinity of deer mice for shrub microhabitats has often been interpreted as a mechanism for reducing predation risk, as deer mice are common prey for many avian and mammalian predators 4

(Clarke 1983, Kotler 1985, Marti et al. 1993). Predator populations on my study area are relatively low (Leslie 1992, P. Stapp, pers. obs.) but deer mice may still prefer shrub cover to minimize perceived risk. Because deer mice are often the most abundant small mammal species where they occur, they have frequently been the focal species for experimental studies of rodent communities (Dueser et al. 1989, Kaufman and Kaufman 1989). Studies of interspecific competition have demonstrated that deer mice are often subordinate to specialized rodent species (Dueser and Hallett 1980, Hallett et al. 1983), but that they quickly colonize areas where competitors are removed (Grant 1971, Redfield et al. 1977, Abramsky et al. 1979, Holbrook 1979, Munger and Brown 1981, Brown and Munger 1985). Grasshopper mice (genus Onychomys) are the only insectivorous rodents in North America, and the largest (30 g) of three species, the northern grasshopper mouse, 0. leucogaster, is the most widespread nocturnal rodent on the Central Plains Experimental Range (Chapter 6). Northern grasshopper mice occur throughout semi-arid and desert grasslands and shrublands from south-central Canada to northern Mexico (Hall 1981) but rarely attain high densities, presumably because their carnivorous habits require large, non-overlapping home ranges (McCarty 1978). Despite the widespread distribution of this species, information on habitat use is limited to broad associations with general vegetation or soil types (e.g. Maxwell and Brown 1968, Kaufman and Fleharty 1974) and virtually nothing is known of its microhabitat affinities. Because grasshopper mice may prey on other rodents (McCarty 1978), understanding the role of grasshopper mice may be useful in interpreting the relative importance of competitive and predatory interactions in the rodent communities. In the following chapters, I describe the results of 4 yr of field studies to address the questions listed above. Chapters 2 and 3 describe autecological studies of deer mice and grasshopper mice, respectively, and Chapters 4 and 5 describe the results of experimental investigations of interactions between the species. In Chapter 6, I outline the methodology for long-term monitoring studies to track populations of rodents and lagomorphs on the Central Plains Experimental Range and describe the first year of results from this research. 5

Habitat selection of Peromyscu.s mtlnicullltus and Onychomys leucogaster Researchers have investigated habitat selection using both theoretical and empirical approaches and small mammals have served as useful model organisms. Habitat-selection theories are based largely on the concept of the ideal free distribution (IFD; Fretwell and Lucas 1970), which predicts that individuals distribute themselves among patches of differing quality according to the expected net gain in resources (and hence, fitness) and intraspecific densities relative to other patches. There has been considerable debate about the value of IFD in experimental studies of animal behavior (Kennedy and Gray 1993, Milinski 1994), but IFD is the foundation of the two prevailing models used to describe habitat selection of free-living small mammals (isodar analysis, Morris 1987a; distribution method, Abramsky et al. 1985, Rosenzweig 1985). Because of the logistical difficulties in measuring habitat use of small secretive animals, habitat selection is assessed indirectly in both approaches by comparing patterns of local abundance among qualitatively distinct habitat patches. The emphasis on depletable resources has permitted the extension of these models from single to multiple species using similar resources because intra- and interspecific competition can be estimated in a similar fashion (Rosenzweig and Abramsky 1986, Abramsky et al. 1991). Rosenzweig and Abramsky and their colleagues have used this approach to study interactions among granivorous rodents (e.g. Abramsky and Pinshow 1989, Ziv et al. 1995) and to generate new theories of how density-dependent habitat selection influences community structure. My studies of habitat use of deer mice and grasshopper mice, in contrast, were largely empirical, in part because the relatively low rodent densities on shortgrass steppe precluded extensive population studies in favor of detailed investigations of microhabitat use. As in earlier field studies of habitat selection by rodents (e.g. Dueser and Shugart 1978, Van Home 1982, Seagle 1985), I also employed multivariate analyses to quantify habitat characteristics and determine how habitat selection changes with scale. However, where previous studies have usually inferred microhabitat affinities by recording characteristics at capture locations, I obtained information on movements and microhabitat use using fluorescent powder tracking 6

(Lemen and Freeman 1985). This technique provides a spatial record of an individual's movements without the bias associated with attraction to bait and trapping... area configuration. In Chapter 2, I demonstrate that deer mice respond to both the overall density and the spatial distribution of shrubs on shortgrass steppe. Patterns of microhabitat use and movement patterns suggested that mice modified their behavior toward shrubs as shrubs became more dense and aggregated. I detected a similar non-linear trend in mouse abundance with increasing shrub canopy cover. I speculate that individual deer mice apparently prefer shrubs to reduce perceived predation risk, although shrubs may also concentrate food resources. At a broader scale, the relationship between abundance and shrub cover may also reflect the scarcity of suitable burrow substrates for deer mice on shortgrass steppe. In contrast, grasshopper mice showed no affinity for shrub microhabitats and were consistently present in most prairie habitats (Chapter 3). Grasshopper mice instead oriented movements toward rodent burrows and disturbances created by pocket gophers (Thomomys talpoides ). Comparison between results from pitfall trapping in different microhabitat types and taxonomic composition of arthropod prey in grasshopper... mouse diets suggested that mice used gopher mounds and burrows because of the higher concentration and availability of insect prey in these microhabitats. Variation in the population density of grasshopper mice was best explained by the abundance of these microhabitats rather than broad-scale, qualitative descriptors of soil or shrub cover type. This result differs from that of Morris ( 1987b ), who found that variation between macrohabitats was a better predictor of the abundance of temperate grassland and forest rodents than microhabitats. One explanation for these differences is that certain resources (e.g., grass, hard mast) differ from mobile insect prey in terms of renewal rates and spatial and temporal predictability. Results from my studies of shrub use by deer mice support this interpretation. I incorporated the microhabitat and macrohabitat variables that were used to study habitat scaling of grasshopper mice (Chapter 3) in a stepwise multiple regression on spatial distribution of deer mice among habitat types. This analysis revealed that deer-mouse 7

density was best explained by variation at the macrohabitat scale alone (F=26.31, P=0.0001, R 2 =0.45) and is consistent with the results of Morris (1987b) for deer mice in other temperate systems. Compared to insects, shrubs therefore may represent a non-depletable resource (i.e., refuge from predation), to which deer mice respond in a relatively fine-grained fashion. Although movements of deer mice were directed toward shrubs and grasshopper mice oriented movements toward mounds and burrows, the movement patterns of individuals of these species were similar in the shrub-grassland study area where both were tracked (Table 1.1). However, as I note in Chapter 3, densities of large shrubs were approximately two-fold higher than those of mounds and burrows. If movements of both species reflected the density of resources alone, then movement indices should differ between the species. How can we reconcile these discrepancies? I used Eberhardt's index to quantify the dispersion of shrubs, soil disturbances, and burrows. Despite differences in the overall density of these microhabitats, the spatial patterning was not different [means (SE) for Eberhardt's index for dispersion of shrubs, disturbances, and burrows were 1.41 (0.04), 1.50 (0.06), and 1.34 (0.05), respectively, for 16 random transects). Thus, one explanation for similarities in movement may be that individual movements reflect the spatial dispersion of resources as well as overall density (Chapter 2). Effects of Onychomys leucogaster on Peromyscus maniculatus and other rodents Studies of small mammals have contributed much to our understanding of the role of interspecific competition in natural communities (Dueser et al. 1989). Over the past decade, however, researchers have recognized that predators also influence population and community dynamics by selectively removing certain taxa or age and sex classes, or by modifying prey behavior (Langland and Jenkins 1987, Brown et al. 1988, Dickman et al. 1991, Dickman 1992, Lima and Dill 1990). The effects of competition and predation traditionally have been considered separately, but in many ecological systems, one or more species may act as both a competitor and predator with species at similar trophic levels. This phenomenon, termed intraguild predation, has been documented in a number of invertebrate 8

communities and may be widespread in assemblages of carnivorous and omnivorous vertebrates as well (Polis et al. 1989). The potential impact of predation by grasshopper mice on deer mice and other small rodents, combined with possible dietary overlap between deer mice and grasshopper mice, was the basis for field experiments of interactions between these species on shortgrass steppe. Furthermore, because of the importance of olfactory communication in interspecific interactions of rodents (Drickamer et al. 1992), especially predator avoidance, I tested the hypothesis that deer mice would avoid odors of grasshopper mice to minimize exposure to this potential predator. I compared patterns of abundance, microhabitat use, and diet of deer mice on four grasshopper-mouse removal sites to those on untreated controls (Chapter 4). Deer mice decreased in number throughout the study on both types of sites, but the decline was greatest on controls, where grasshopper-mouse numbers increased. I detected no microhabitat shifts on removal sites, but deer mice increased their use of shrubs on controls when grasshopper mice became abundant. Because deer-mouse diets did not differ between control and remova1 sites during the experiment and because mice increased their use of microhabitats typically not used by grasshopper mice, I concluded that grasshopper mice affected deer mice through predation or interference rather than resource competition. Increases in the numbers of granivorous rodents on removals support this conclusion, and suggest that, when abundant, grasshopper mice may impact the structure of local rodent assemblages. A major criticism of field experiments in community ecology has been the lack of implementation of rigorous study designs that permit accurate conclusions (Hurlbert 1984, Galindo and Krebs 1986, Dueser et al. 1989). Although my experiment was relatively short in duration (3 months), overall I employed a higher level of replication, larger plot sizes, and more frequent monitoring and maintenance of removal effects than previous small-mammal studies (Dueser et al. 1989). Despite of these efforts, natural fluctuations in the deer-mouse abundance confounded my results. However, the area-wide decline in deer mice also led me to speculate that the relative importance of interspecific interactions between deer mice and 9

grasshopper mice may vary depending on background levels of resource abundance and other external factors that determine successful reproduction of deer mice on a broader scale. For example, in areas of co-occurrence, predation by grasshopper mice during normal years may be opportunistic and focused primarily on juveniles or litters, especially given the affinity of grasshopper mice for rodent burrows. During years of normal or high reproduction, animals lost to predators may be replaced by recruitment from surrounding areas. My experimental plots were intentionally established adjacent to floodplain vegetation, where deer mice consistently reached higher densities, had a more-even sex ratio, a higher proportion of the population reproductive, earlier and more consistent production of juveniles, higher apparent survival rates, and smaller home ranges than in areas of mixed shrub-grassland (P. Stapp, unpublished data). I therefore might not have detected a numerical response of deer mice to changes in grasshopper-mouse numbers had I conducted my study when deer mice were abundant. Furthermore, a decrease in the abundance of insects that reduced food availability for grasshopper mice may cause grasshopper mice to seek rodent prey more actively. For example, predation on other rodents may occur more frequently during winter, when arthropods are less abundant. These results underscore the importance of studying resource availability and the behavior of individuals to understand the mechanisms underlying community patterns. Although my removal experiment suggests that grasshopper mice affect deer mice in areas where they co-occur, my odor-response experiment provided no evidence that this relationship is mediated by olfactory cues (Chapter 5). When presented with odors of grasshopper mice, harvest mice, and clean cotton, deer mice showed no avoidance of grasshopper-mouse odors. This result was somewhat surprising from previous research demonstrating that rodents generally avoid predator odors, but a detailed review of the studies of odor response of Peromyscus species revealed little evidence that these mice respond to heterospecific odors. 10

Insights and unanswered questions Taken with earlier small-mammal studies conducted on the Central Plains Experimental Range, my results provide some general insights into what shapes rodent communities of shortgrass steppe (Fig. 1.1). For species such as deer mice, western harvest mice (Reithrodontomy megalotis), and also prairie voles (Microtus ochrogaster), local abundance is probably determined largely by the abundance of vegetation cover, which provides both food and protection from predators. Granivorous rodents such as kangaroo rats (Dipodomys ordii) and locally-rare pocket mice (Chaetodipus hispidus, Perognathus spp.) respond primarily to soil type through its effect on the production and availability of palatable seeds of annual forbs and intermediate-height grasses. The distribution of grasshopper mice also reflects edaphic conditions, but apparently via the effects of soil friability on other, fossorial rodents and on the consequences of these animals' activities on the availability of arthropod prey. In areas of shortgrass steppe where habitat conditions permit coexistence, grasshopper mice may modify the behavior and population dynamics of deer mice and other rodents. However, the role of an opportunistic predator such as the grasshopper mouse, and of other species interactions, ultimately must be interpreted in the context of overall resource availability, which likely varies with fluctuations in abiotic conditions and the resulting effects on productivity. Human activities such as intensive cattle-grazing and plowing that reduce the structure and diversity of vegetation may have direct effects on species that require seeds or cover, but may also improve conditions for grasshopper mice and hence, influence community structure indirectly. My findings suggest at least three questions that merit further exploration. First, how general are the apparent non-linear relationships between resource distributions, habitat use, and abundance described for deer mice in Chapter 2? These patterns suggest that our assessment of the importance of habitat features may differ depending on the circumstances (e.g., resource abundance, population density) under which habitat selection is measured, which may complicate elucidation of wildlife-habitat relationships and, as a consequence, 11

conservation and management efforts. Second, to what degree can differences in patterns of habitat use between organisms at different trophic levels be explained by intrinsic differences in the spatial and temporal predictability and renewal rates of their critical resources? In the context of other studies of habitat selection, my studies suggest that there may be basic differences in scaling of habitat use between granivorous, herbivorous, and carnivorous rodents that reflect differences in the distribution and availability of resources. Finally, traditional studies of predator-prey relationships are often concerned with how predators impact prey populations through direct mortality, but the interactions between omnivorous and carnivorous species may depend indirectly on the abundance of non-shared resources. What are the ecological effects of opportunistic predation and under what conditions can they affect the evolution of local assemblages? Literature Cited Abramsky, Z., and B. Pinshow. 1989. Changes in foraging effort in gerbil species correlate with habitat type and intra- and interspecific activity. Oikos 56:43-53. Abramsky, Z., M.I. Dyer, and P.D. Harrison. 1979. Competition among small mammals in experimentally perturbed areas of the shortgrass prairie. Ecology, 60:530-536. Abramsky, Z., M.L. Rosenzweig, and S. Brand. 1985. Habitat selection of Israel desert rodents: a comparison of a traditional and a new method of analysis. Oikos 45:79-88. Abramsky, Z., M.L. Rosenzweig, and B. Pinshow. 1991. The shape of a gerbil isocline measured using principles of optimal habitat selection. Ecology 72:329-340. Andrewartha, H. G., and L.C. Birch. 1954. The distribution and abundance of animals. University of Chicago Press, Chicago. Baker, R.H. 1968. Habitats and distribution. Pp. 98-121 in: Biology of Peromyscus (Rodentia) (J.A. King, ed.). Special Publication No.2, American Society of Mammalogists, Stillwater, OK. Batzli, G.O., and C. Lesieutre. 1995. Community organization of arvicoline rodents in northern Alaska. Oikos 72:88-98. Brown, J.H., and B.A. Harney. 1993. Population and community ecology of heteromyid rodents in temperate habitats. Pp. 618-651 in Biology of the Heteromyidae (H.H. Genoways and J.H. Brown, eds.). Special Publication No. 10, American Society of Mammalogists. 12

Brown, J.H., and J.C. Munger. 1985. Experimental manipulation of a desert rodent community: food addition and species removal. Ecology 66:1545-1563. Brown, J.S., B.P. Kotler, R.J. Smith, and W.O. Wirtz n. 1988. The effects of owl predation on the foraging behavior ofheteromyid rodents. Oecologia 76:408-415. Clarke, J.A. 1983. Moonlight's influence on predator/prey interactions between short-eared owls (Asio flammeus) and deer mice (Peromyscus maniculatus). Behavioral Ecology and Sociobiology 13:205-209. Dickman, C. R. 1992. Predation and habitat shift in the house mouse Mus domesticus. Ecology 73:313-322. Dickman, C.R., M. Predavec, and A.J. Lynam. 1991. Differential predation of size and sex classes of mice by the barn owl, Tyto alba. Oikos 62:67-76. Drickamer, L.C., D.O. Mikesic, and K.S. Shaffer. 1992. Use of odor baits in traps to test reactions to intra- and interspecific chemical cues in house mice living in outdoor exclosures. Journal of Chemical Ecology 18:2223-2250. Dueser, R.D., J.H. Porter, and J.L. Dooley, Jr. 1989. Direct tests for competition in North American rodent communities: synthesis and prognosis. Pages 105-125 in: Patterns in the structure of mammalian communities (D.W. Morris, Z. Abramsky, B.J. Fox, and M.R.Willig, eds.). Texas Tech University Press, Lubbock, Texas. Dueser, R.D., and H.H. Shugart, Jr. 1978. Microhabitats in a forest-floor small mammal fauna. Ecology 59:89-98. French, N.R., W.E. Grant, W. Grodzinski, and D.M. Swift. 1976. Small mammal energetics in grassland ecosystems. Ecological Monographs 46:210-220. Fretwell, S.D., and H.L. Lucas, Jr. 1970. On territorial behavior and other factors influencing habitat distribution in birds. I. Theoretical development. Acta Biotheoretica 19:16-36. Galindo, C., and C.J. Krebs. 1986. Evidence for competition in small rodents. Oikos 46:116-118. Grant, P.R. 1971. Experimental studies of competitive interaction in a two-species system. m. Microtus and Peromyscus species in enclosures. Journal of Animal Ecology 40:323-350. Grant, W.E., and E.C. Birney. 1979. Small mammal community structure in North American grasslands. Journal of Mammalogy 60:23-36. Grant, W.E., and N.R. French. 1980. Evaluation of the role of small mammals in grassland ecosystems: a modelling approach. Ecological Modelling 8:15-37. 13

Hall, E.R. 1981. Mammals of North America. 2nd edition. Vol. 2. John Wiley and Sons, New York, New York. Hallett, J.G., M.A. O'Connell, and R.L. Honeycutt. 1983. Competition and habitat selection: test of a theory using small mammals. Oikos 40: 175-181. Holbrook, S.J. 1978. Habitat relationships and coexistence of four sympatric species of Peromyscus in northwestern New Mexico. Journal ofmammalogy 51:18-26. Holbrook, S.J. 1979. Habitat utilization, competitive interactions, and coexistence of three species of cricetine rodents in east-central Arizona. Ecology 60:758-769. Hurlbert, S.H. 1984. Pseudoreplication and the design of field experiments. Ecological Monographs 54:187-211. Kaufman, D.W., and E.D. Fleharty. 1974. Habitat selection by nine species of rodents in north-central Kansas. Southwestern Naturalist, 18:443-452. Kaufman, D.W., and G.A. Kaufman. 1989. Population biology. Pp. 233-270 in: Advances in the study of Peromyscus (Rodentia) (G.L. Kirkland, Jr., and J.N. Layne, eds.). Texas Tech University Press, Lubbock, Texas. Kennedy, M. and R.D. Gray. 1993. Can ecological theory predict the distribution of foraging animals? A critical analysis of experiments on the Ideal Free Distribution. Oikos 68:158-166. Kotler, B.P. 1984. Risk of predation and the structure of desert rodent communities. Ecology 65:689-701. Kotler, B.P. 1985. Owl predation on desert rodents which differ in morphology and behavior. Journal of Mammalogy 66:824-828. Kotler, B.P., and J.S. Brown. 1988. Environmental heterogeneity and the coexistence of desert rodents. Annual Review of Ecology and Systematics 19:281-307. Kotler, B.P., J.S. Brown, and W.A. Mitchell. 1994. The role of predation in shaping the behaviour, morphology, and community organisation of desert rodents. Australian Journal of Zoology 42:449-466. Lauenroth, W.K., and D.O. Milchunas. 1991. Short-grass steppe. Pp. 183-226 in: Natural Grasslands, Volume 1 (R.T. Coupland, ed.), Ecosystems of the World 8A, Elsevier, Amsterdam. Lemen, C.A. and P.W. Freeman. 1985. Tracking mammals with fluorescent powder pigments: a new technique. Journal ofmammalogy 66:134-136. Leslie, D.O. 1992. Population status, habitat, and nest-site characteristics of the raptor community on the Pawnee National Grasslands. M.Sc. thesis, Colorado State 14

University, Fort Collins, CO. Lima, S.L., and L.M. Dill. 1990. Behavioral decisions made under the risk of predation: a review and prospectus. Canadian Journal of Zoology 68:619-640. Longland, W.S., and S.H. Jenkins. 1987. Sex and age affect vulnerability of desert rodents to owl predation. Journal of Mammalogy 68:746-754. Marti, C.D., K. Steenhof, M.N. Kochert, and J.F. Marks. 1993. Community trophic structure: the roles of diet, body size, and activity time in vertebrate predators. Oikos 67:6-18. Maxwell, M.H. and L.N. Brown. 1968. Ecological distribution of rodents on the High Plains of eastern Wyoming. Southwestern Naturalist, 13:143-158. McCarty, R. 1978. Onychomys leucogaster. Mammalian Species 87:1-6. Milinski, M. 1994. Ideal free theory predicts more than input matching - a critique of Kennedy and Gray's review. Oikos 71:163-166. Morris, D.W. 1987a. Spatial scale and the cost of density-dependent habitat selection. Evolutionary Ecology 1:379-388. Morris, D.W. 1987b. Ecological scale and habitat use. Ecology, 68:362-369. Munger, J.C., and J.H. Brown. 1981. Competition in desert rodents: an experiment with semipermeable exclosures. Science 211:510-512. O'Farrell, M.J. 1980. Spatial relationships of rodents in a sagebrush community. Journal of Mammalogy 61:589-605. Polis, G.A, C.A. Myers, and R.D. Holt. 1989. The evolution and ecology of intraguild predation: competitors that eat each other. Annual Review of Ecology and Systematics 20:297-330. Pulliam, H.R. 1988. Sources, sinks, and population regulation. American Naturalist 132:652-661. Redfield, J.A., C.J. Krebs, and M.J. Taitt. 1977. Competition between Peromyscus maniculatus and Microtus townsendii in grasslands of coastal British Columbia. Journal of Animal Ecology 46:607-616. Reichman, O.J., and M.J. Price. 1993. Ecological aspects of heteromyid foraging. Pp. 539-574 in: Biology of the Heteromyidae (H.H. Genoways and J.H. Brown, eds.). Special Publication No. 10, American Society of Mammalogists. Rose, R.K., and B.C. Birney. 1985. Community ecology. Pp. 310-339 in: Biology of New World Microtus (RH. Tamarin, ed.), Special Publication No. 8, American Society of Mammalogists. 15

Rosenzweig, M.L., and Z. Abramsky. 1985. Detecting density-dependent habitat selection. American Naturalist 126:405-417. Rosenzweig, M.L., and Z. Abramsky. 1986. Centrifugal community organization. Oikos 46:339-348. Seagle, S.W. 1985. Patterns of small mammal microhabitat utilization in cedar glade and deciduous forest habitats. Journal ofmammalogy 66:22-35. Sib, A., P. Crowley, M. McPeek, J. Petranka, and K. Strohmeier. 1985. Predation, competition, and prey communities: a review of field experiments. Annual Review of Ecology and Systematics 16:269-311. Stephens, D.W., and J.R. Krebs. 1986. Foraging theory. Princeton University Press, Princeton, NJ. Thompson, S.D. 1982. Structure and species composition of desert heteromyid rodent species assemblages: effects of a simple habitat manipulation. Ecology 63:1313-1321. Travers, S.E., D.W. Kaufman, and G.A. Kaufman. 1988. Differential use of experimental habitat patches by foraging Peromyscus maniculatus on dark and bright nights. Journal of Mammalogy 69:869-872. Van Home, B. 1982. Niches of adult and juvenile deer mice (Peromyscus maniculatus) in seral stages of coniferous forest. Ecology 63:992-1003. Van Home, B. 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47:893-901. Wiens, J.A. 1989. The ecology of bird communities. Volume 1. Foundations and patterns. Cambridge University Press, Cambridge. Ziv, Y., B.P. Kotler, Z. Abramsky, and M.L. Rosenzweig. 1995. Foraging efficiencies of competing rodents: why do gerbils exhibit shared-preference habitat selection? Oikos 73:260-268. 16

Table 1.1 Movement indices of powder trails from deer mice tracked during December 1992 and August 1993, and grasshopper mice tracked during January and July 1994 on the shrubgrassland trapping area. Values are means for n paths, with standard errors in parentheses. n Mean vector length Fractal dimension Deer mouse 8 0.83 (0.03) 1.16 (0.03) Grasshopper mouse 12 0.79 (0.01) 1.17 (0.02) 17

SOIL TYPE I TEXTURE ------ ~/~~ GRASS SHRUBS CACTUS BARE SOIL. I, ~ : ~ :., \, I I t I I, I f,.,. ""'/ GOPHER MOUNDS \..... I \ ~ f 00 [Seeds]... lcoverl raithioj)o:a&], ::... ',........... Reithrodontomys Dipodomys Peromyscus.. Onychomys Fig. 1.1. Flow diagram depicting ecological relationships among habitat characteristics (capitalized), resources (in boxes), and the four most common nocturnal rodents (italics) in shortgrass steppe. Interactions represented by stippled arrows were not investigated directly during my research.

CHAPTER2 RESPONSE OF DEER MICE (PEROMYSCUS MANICUIATUS) TO SHRUBS: LINKING SMALL-SCALE MOVEMENTS AND THE SPATIAL DISTRIBUTION OF INDIVIDUALS ABSTRACT The distribution of individuals and populations may reflect the abundance and spatial distribution of resources across a range of scales but there have been relatively few attempts to link insights from studies of these different phenomena, especially for wide-ranging vertebrates. I live-trapped and tracked deer mice (Peromyscus maniculatus) across a gradient of shrub cover on shortgrass steppe in north-central Colorado to estimate population size and quantify patterns of movement and microhabitat use. Mice appeared to prefer shrub microhabitats, especially in areas where shrubs were less numerous, and to orient their movements toward shrubs and shrub patches. Population density also increased with increasing density and aggregation of shrubs. Furthermore, thresholds in the selective versus random use of shrub microhabitats, movement patterns, and population density occurred over a narrow range of shrub cover where shrubs were most aggregated, underscoring the importance of both the density and dispersion of shrubs. Relationships between shrub cover and movement parameters and abundance suggested that mice accumulated in areas where their movements were most tortuous. Information on movements of individuals therefore can produce testable predictions about patterns of local abundance and may provide insights into the relationship between space use and population size. 19

INTRODUCTION Ecologists have generally approached the study of natural populations from two directions (Hassell and May 1985). Behavioral ecologists typically seek adaptive explanations for the responses of individuals to resources, conspecifics and predators, with the objective of discerning how these behaviors ultimately contribute to survival and reproductive success. Population ecologists, on the other hand, are concerned primarily with demographic processes such as mortality, fecundity and dispersal, and the consequences of these factors for the distribution and abundance of individuals. Behavioral and population phenomena are usually addressed at different spatial and temporal scales, but the study of animal movement is central to both approaches (Stenseth and Lidicker 1991 ). Indeed, several authors (e.g. Crist and Wiens 1995, Wiens et al. 1993, Johnson et al. 1992, Turchin 1991) have argued that the spatial structure of populations emerges from the collective behavior of individuals interacting with landscape features, and that knowledge of the mechanisms underlying the movements of individuals will improve our understanding of patterns of spatial distribution and abundance. Researchers have attempted to link individual behavior and population phenomena using a variety of theoretical approaches, including models based on game theory (e.g. Goss Custard et al. 1995), differential equations of predator-prey and population dynamics (Kareiva and Odell 1987, Hassell and May 1985), individual-based movements (reviewed by Johnson et al. 1992), and diffusion or random walks (e.g. Crist and Wiens 1995, Gautestad and Mysterud 1993, Benhamou and Bovet 1989). Turchin (1991) reviewed diffusion and random-walk approaches and proposed a mathematical model to calculate the spatial distribution of foragers from empirically-derived parameters of individual movements. Wiens and his co-workers (Crist and Wiens 1995, Wiens et al. 1995, Crist et al. 1992, Johnson et al. 1992, Wiens and Milne 1989) have advocated the use of experimental model systems, in which one records the behavior of small organisms, usually insects, in artificial or manipulated "micro landscapes". The goal of this approach is to extrapolate pattern-process 20

relationships from rigorous, small-scale experiments to ecological systems at broader scales. These efforts have contributed to our understanding of the mechanistic basis of animal movement, but applications of these insights to natural systems involving wide-ranging vertebrate organisms have remained elusive. In this paper, I describe how patterns of abundance of a small mammal reflect responses to the spatial distribution of habitat components that are similar to those observed in movements of individuals within their home ranges. I recorded the nocturnal movements of deer mice [Peromyscus maniculatus nebracensis (Coues)] with respect to woody shrubs, and compared patterns of abundance to the distribution of shrub cover on shortgrass steppe, an area of semiarid grasslands in the central United States. Shortgrass steppe is dominated by short, perennial grasses and has little vertical structure except in low-lying areas, where soil texture permits establishment of large shrubs (Lauenroth and Milchunas 1991). Small rodents such as deer mice are more abundant in shrub-dominated areas than in open grasslands (Chapter 6; Lindquist et al. 1995), reflecting the importance of vegetative cover for quadrupedal rodents in arid and semiarid regions of North America (Kotler and Brown 1988). Deer mice attain a wide range of densities across gradients of shrub cover on my study area in north-central Colorado, which allowed me to document habitat selection at microhabitat and macrohabitat scales (Rosenzweig 1989). My study addresses four questions. First, do deer mice prefer shrub microhabitats in shortgrass steppe? Second, how do the movements of individual deer mice reflect their response to shrub cover? Third, what is the relationship between the relative abundance of mice and the abundance and spatial distribution of shrubs? Last, can we predict the spatial distribution of deer mice along gradients of shrub cover from the movement characteristics of individuals? MATERIALS AND METHODS My study area was the Central Plains Experimental Range, located ca. 60 km northeast of Fort Collins, Colorado, USA. Research at the 6200-ha site is coordinated by the 21

USDA Agricultural Research Service and the Shortgrass Steppe Long-Term Ecological Research project (National Science Foundation). The climate is semi-arid: mean monthly temperatures range from -5 C in January to 22 C in July, and ca. 70% of the 321 mm of annual precipitation falls during late spring and summer (Milchunas and Lauenroth 1995). Upland vegetation is short in stature and dominated by two perennial bunchgrasses [Bouteloua gracilis (H.B.K.) Lag. and Buchloe dactyloides (Nutt.) Engelm.], with a mixture of small shrubs and forbs [Artemisia frigida Willd., Eriogonum effusum Nutt., Sphaeralcea coccinea (Pursh) Rydb.], and prickly-pear cactus (Opuntia polycantha Haw.). Low-lying areas have abundant large shrubs [primarily four-wing saltbush, Atriplex canescens (Pursh) Nutt.] and a variety of half-shrubs [A. frigida, E. effusum, Gutierrezia sarothrae (Pursh) Britt.and Rushy, Ceratoides lanata (Pursh) J.T. Howell, Chrysothamnus nauseosus (Pall.) Britt.] and mid-grasses (Pascopyron smithii (Rydb.) A. LOve, Stipa comata Trin. and Rupr., Aristida longiseta Steud., Sitanion hystrix (Nutt.) J.G. Smith, Oryzopsis hymenoides (Roem.and Schult.) Ricker]. Small soapweed (Yucca glauca) occurs primarily on ridges and on sandy soils. A broad floodplain associated with two ephemeral creeks is on loamy soils with thick grass (B. gracilis and P. smithii) and dense saltbush. Upland sites are grazed by cattle in summer and fall, whereas shrub areas are mostly grazed in winter and spring. Gradients in shrub density occur on the study area, with high densities of saltbush on the floodplain, intermediate shrub densities on adjacent coarsely-textured soils, and no large shrubs on upland prairie. I estimated relative densities of deer mice by live-trapping on sites with varying amounts of shrub cover. Mice were trapped on three 1.82-ha rectangular grids in winter 1992193 (20 December 1992-31 January 1993) and summer 1993 (16-28 July), nine 1.54-ha circular webs (Buckland et al. 1993) in early summer 1994 (7-15 June) and 25 0.32-ha rectangular plots in early summer 1995 (27 May- 3 July; Table 2.1). Relative densities were calculated as the number of individuals captured on each site, divided by the effective trapping area. The effective trapping area (Table 2.1) was the area bounded by the traps, plus a 18-m boundary strip to adjust for movement of mice onto the trapping areas. 22